C. BRENT BARGERON, RICHARD C. BENSON, ROBB W. NEWMAN, A. NORMAN JETTE, and TERRY E. PHILLIPS OXIDATION MECHANISMS OF HAFNIUM CARBIDE AND HAFNIUM DIBORIDE IN THE TEMPERATURE RANGE 1400 TO 2100°C Two ultra-high-temperature materials, hafnium carbide and hafnium diboride, were oxidized in the temperature range 1400 to 2100°C. The two materials oxidized in distinctly different ways. The carbide formed a three-layer system consisting of a layer of residual carbide, a layer of reduced (partially oxidized) hafnium oxide containing carbon, and a layer of fully oxidized hafnium dioxide. The diboride oxidized into only two layers. For the diboride system, the outer layer, mainly hafnium dioxide, contained several intriguing physical structures. INTRODUCTION Materials that can provide protection at temperatures aspects of research that has been performed over the past above l700°C in an oxidative environment are needed for four or five years.2-5 important applications. To be usefully employed as a turbine blade coating, for example, a substance would EXPERIMENTAL METHODS need to withstand many excursions from normal ambient The experimental arrangement for the oxidation pro­ conditions into the high-temperature regime and back cess has been described in detail previously.2-4 An induc­ again without cracking, spalling, or ablating. Other ap­ tion furnace consisting of two concentric zirconia tubes plications, such as a combustion chamber liner, might with a graphite susceptor between them was used to heat require only one high-temperature exposure. Not only do the specimen. (A susceptor is the heating element in an the chemical properties need to be considered, but the induction furnace.) The specimen temperature was mea­ physical, microscopic structure of a candidate material sured with an optical pyrometer through a sapphire win­ can also determine how well it will function under ex­ dow. Gas flow through the furnace of about 20 cm/s was treme conditions. To illustrate, a substance fabricated by controlled with a system of solenoid valves that main­ hot-pressing a powder will generally have micrometer­ tained a flow of argon until the desired temperature was size pores throughout, which allow hot oxidative gases reached, at which time oxygen, at a partial pressure of 55 to diffuse at a significantly higher rate than in a material torr, was added to the furnace atmosphere. After a spec­ of the same chemical composition made by chemical ified oxidation time, the atmosphere was switched back vapor deposition (CYD). Carefully manufactured CYD to pure argon, the furnace was turned off, and the spec­ films do not have pores or cracks, so hot gases must imen was allowed to cool to room temperature. diffuse through the lattice of the bulk material, a much The specimen was then removed from the zirconia slower process than diffusion through pores or openings. tube, photographed at several magnifications, embedded The oxidation mechanisms of CYD films of two impor­ in epoxy, cut to reveal a cross section of the oxidized film, tant high-temperature materials are discussed in this and finally polished using a series of successively finer article. Hafnium carbide and hafnium diboride melt at diamond grit. Normally, samples were sectioned by cut­ 3950 and 3250°C, respectively. In an oxygen atmosphere, ting perpendicular to the surface. Occasionally, however, both form hafnium dioxide, which melts at about a specimen was cut at a very small angle to the surface 2900°C. In addition to at least one solid oxide, both to expose broader sections of oxide and residual material substances produce other oxidation products. Hafnium for X-ray diffraction measurements. Thickness measure­ carbide yields either CO or CO2, and hafnium diboride ments of the various layers were made with a calibrated produces boric oxide, which becomes a liquid at 450°C Olympus metallurgical microscope, and cross-sectional and vaporizes at 1860°C. Obviously the presence of photographs were taken. The polished cross sections were gases and liquids in a protective film can have significant also examined by X-ray microanalysis in a scanning ramifications. Further, a structural phase transition that electron microscope (SEM). occurs in hafnium dioxide at about 1700°C should not To characterize certain layers, electrical resistance was be overlooked. This transition entails an accompanying measured as a function of temperature. Leads were at­ 3.4% volume changel that can result in cracking or spall­ tached successively to the polished surfaces of the indi­ ing. The work presented here is a selection of various vidual layers, and the specimen was then placed in a f ohns Hopkins APL Technical Digest, Vo lume 14, Number 1 (/993) 29 C. B. Bargeron et ai. cryostat, where the resistance was measured as the tem­ observe hollow channels in the protrusions. Some of the perature was lowered to near 10K and then returned to stovepipes have a height almost equal to the film thick­ room temperature. The hardness numbers were deter­ ness. The formation of these structures is apparently an mined for the oxidized hafnium carbide film using stan­ intricate process that includes the transport and deposi­ dard procedures employing a Knoop indenter. Polished tion of hafnium dioxide by liquid or gaseous B20 3. Figure surfaces were also employed for the latter measurements. 4 is a stereopair of photographs of one of the stovepipes, revealing in dramatic fashion not only the hollow center RESULTS of the protrusion but also a large channel going into the A photograph of the hafnium diboride film before ox­ oxidized layer at the left of the structure's base. idation is shown in Figure l. The polycrystalline film has Two examples of oxide growth morphology formed at well-formed crystallites with flat faces and well-defined 1900°C are shown in Figure 5. The hafnium diboride edges. The overall integrity of the coating is good; only exposed for 100 s (Fig. 5A) has an interesting growth an occasional crack is present. On the basis of the exam­ pattern in which the oxide has periodically separated into inations of cross sections revealing the film/substrate in­ thin laminae. The film exposed to oxygen for 300 s (Fig. terface, the attachment of the coating to the substrate is 5B) provides an example of the formation of large voids. believed to be excellent. Similar observations were made In both examples, the separation of oxide and the remain­ for the hafnium carbide films. ing diboride is apparent. Figure 2 shows a cross section of the film after expo­ Examination of cross sections of hafnium carbide re­ sure to oxygen for 1800 s at 1520°C. At this temperature, vealed a phenomenon different from that of hafnium a relatively compact oxide forms. Cracks in the oxide diboride in that an additional layer had formed between between many of the grains were probably formed during the residual carbide and the white outer layer of oxide. cooling or post-oxidation handling. Oxide formed at this In the light microscope, the new layer appeared gray (Fig. temperature appears to adhere to the residual diboride. At 6). In the SEM, the interlayer was observed to be compact higher oxidation temperatures, however, separation be­ and fine grained compared with the porous outer oxygen tween the oxide and diboride is common and is probably layer shown in Figure 7. This new layer was very intrigu­ caused by different temperature coefficients of expansion ing from the point of view of protecting the substrate, for the oxide and diboride. since compact layers are expected to allow much slower Near the boiling point of boric oxide C::d860°C), the transport of gases than porous materials. To characterize oxidized hafnium diboride develops some interesting the interlayer, further analysis was performed. morphology. Figure 3 is a stereopair of photographs of Figures 8A, 8B , and 8C show X-ray microanalysis the oxide after exposure to oxygen for 540 s at 1850°C. spectra of the residual carbide, the interlayer, and the Many blunt posts protrude from the surface, a few of outer oxide layer, respectively. As noted earlier,3 the big which have no cover over the top, suggesting that they change in the oxygen-to-hafnium peak ratio occurs be­ are hollow, resembling stovepipes. In stereo, one may tween the residual carbide and the interlayer. In other Epoxy Oxide Diboride 20 ttm Figure 2. Cross section of film after exposure to 55 torr of oxygen for 1800 s at 1520°C. The oxide is compact but has some cracks I 40 ttm I between the grains. The cracks probably formed during cooling or Figure 1. Surface of hafnium diboride film showing crystallites post-experimental handling. Oxide formed at this temperature with flat faces and well-defined edges. appears to adhere to the diboride in most places. 30 Johns Hopkins APL Technical Digest, Vo lume 14, Number 1 (1 993) Oxidation Mechanisms of Hafnium Carbide and Hafnium Diboride Epoxy Figure 3. A stereopair of photographs of the cross section of a film oxidized in 55 torr of oxygen for 540 s at 1850°C, showing unusual protrusions from the f· oxide surface, Viewing in stereo reveals 'c·:-: , .. ,.s'O,C' " Oxide that these structures are hollow (see also Fig . 4), resembling stovepipes, with a well-formed central channel. In fact, many channels occur throughout the Gap films oxidized at this temperature, sev­ eral of which can be observed in this stereopair. Note also that the oxide is Diboride separated from the residual diboride. == Substrate 200 JLm Figure 4. A higher-magnification stereopair of photographs of the stove­ pipe on the left in Figure 3. Viewing in stereo reveals a hollow, smooth-walled central tube leading out of the oxide. The cracking in the wall is probably due to post-experimental handling.